Date: Fri, 12 Sep 1997 05:53:54 -0400 (EDT) From: Peter Dufault <dufault@hda.com> To: hackers@freebsd.org Subject: Re: Realtime Programming Under FreeBSD? Message-ID: <199709120953.FAA26739@hda.hda.com> In-Reply-To: <199709120402.VAA13813@usr03.primenet.com> from Terry Lambert at "Sep 12, 97 04:02:56 am"
next in thread | previous in thread | raw e-mail | index | archive | help
> Yes. This was clarified after the posting to which you are responding; > you are correct that he meant completely orthoganal scheduling. > > This will never fly, of course, since it means that the overall > system performance will be degraded in the general case. Yes. But you may be perfectly willing to live with a performance hit in some cases. And the reason you must do this is the orthogonal kernel has work to do - it may be giving time slices to user threads, etc, and without thinking it over more carefully I'm not sure the real time part can give up its piece of the duty cycle when it has nothing to do. Why is this useful? Some reasons include: 1. You're testing something on your 200Mhz Pentium that will later run standalone on your 66Mhz 486; 2. You're moving toward better real time, comparable to the push down of the giant lock in SMP; 3. You're working toward dedicated resource real time using a multiprocessor, and you're essentially simulating the other processors with time multiplexing. Ad-hoc approaches boil down to this, anyway. Peter -- Peter Dufault (dufault@hda.com) Realtime development, Machine control, HD Associates, Inc. Safety critical systems, Agency approval
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199709120953.FAA26739>